The Clockwork Game Design Podcast: Episode 12 – Verificationism and Goals

cgdplogo_mediumIn this episode, I really talk about three things:

  1. I wanted to address some comments I got about last episode and whether I really agreed with much of what Charles Pratt said, as I seemed to to some listeners
  2. The idea of “verificationism” which I was introduced to by Sam Harris’ latest podcast episode, and
  3. Goals – what makes a good goal, and what should we avoid in goal design?

Some stuff I referenced in the podcast:

Sam Harris’ latest podcast episode, which is absolutely recommended listening for my listeners

My book, which has a lot more detail on good goal design

Wikipedia article on verificationism

My 3 Minute Game Design episode, “How Games Work” which touches on goals (other episodes in the series do as well)

Thanks for listening, and as always, you can support the show by visiting my Patreon page. Patrons will soon gain access to a new top secret prototype I’m working on, by the way!

  • Jereshroom

    I get that a binary goal can’t *literally* be “get the highest score you can”, because you are guaranteed to win.
    But I still like higher-score-is-better systems. When it comes down to it, that’s how all games work — chess players want to win as many games “as they can”. I want to beat a solitaire game in “as few tries as possible”. And IRL I want to be as happy “as I can”.
    As long as you consider a score of 2 to be exactly half as good as a score of 4 (and so on), you don’t run into any conflicts of interest — you just figure out what score you will get on average if you do/don’t make a certain decision. Eg. 60% chance to beat 3 bosses and 30% chance to get to beat 1 boss would be a worse option that being guaranteed to beat 2 bosses.
    If your argument is just that art without a binary goal is not a game, then I think it would be best to change your definition of “goal” slightly. Because your current definition is leaving out too many game-like things.

  • I’m not talking about real life, series of chess matches, or anything else beyond “inside a match of chess, what is your goal”. The goal cannot be “capture as many pieces as you can” in that context.

    I can maybe see another word for goals in games that makes it distinct from the goals in contests, though.

  • Jereshroom

    So is there anything wrong with more-is-better systems? Is your only issue with them the fact that they aren’t games? Because more-is-better systems work incredibly well highlighting and working with core mechanisms. For instance, a “climb as high as you can” game has a clear core mechanic of climbing upwards.

    Also: Would you call more-is-better systems contests, or toys?

  • Usually a “climb as high as you can” type of thing would be a contest, not a game. There’s nothing wrong with them, but they do become problematic when you have basically the structure of a game (decision-making) but missing the kind of goal that a game needs.

  • Jereshroom

    Because you don’t know the comparative value of scores, such as whether to risk a low score to get a high score, or settle for a decent score, right?What if you give the players something of value based on their score? Poker, for instance, is rarely played as what you would count as a game. But it still functions well because a 50% chance of $2 is worth the same as $1. So there is only one logical way to calculate the value of each decision.

  • Jimmy Mac

    I don’t think many players play for “the highest score I can get”. I think most players are playing for “a higher score than the current high scorer”. A player may say, “Well, at least I scored higher than my previous best!” as a way to track their progress, but I don’t think that’s as satisfying as beating another player. I agree that it can end up feeling like a contest, not a game, maybe because players aren’t interacting.

  • Jimmy Mac

    I’m surprised you don’t agree with logical positivism/verificationism. I was talking with you in the comments on episode 10 ( where you said:

    >So far everything else in the universe we’ve taken a hard look at has had objectively better and worse answers, so to me it seems reasonable to believe that it’s likely to be the case for design as well.

    My take on your stance is that 1) there are empirical or at least logically verifiable principles of game design and 2) empirical/verifiable principles are preferable. Is that right? How do you see what you think as different from logical positivism/verificationism?

  • Jereshroom

    But most people have essentially no chance of getting on the high score board, so they really just play to get “a good score” or something like that — which is a really messy kind of goal.

  • Of course everyone agrees with 2. The issue with verificationists is that it’s not that “empirical/provable statements are PREFERABLE”, it’s that they’re the only things at all worth saying. Whereas a lot of the more specific game design findings, like my work, a lot of it is sort of reasoned proposals and ways of looking at aesthetics. I think a lot of verificationists would reject a lot of my work because I can’t “prove” it in the way that we can prove some mathematical equation. Verificationists basically reject soft sciences like philosophy and aesthetics.

  • Jesse Fortner

    What do you mean when you say they have the decision-making structure of a game? What degree of decision-making is acceptable in a contest?

  • Well, no decision making is allowed in a contest, formally speaking.

  • Jesse Fortner

    So if a game requires a goal, and goals must be binary; and interactive systems with measurement but not a binary goal is a contest, which should not have decision-making. If that is true, then an interactive system which has both a non-binary “goal” and decision-making does not fit neatly into either form, and thus should be conceptually discarded. Is this a correct application of the idea of forms?

  • I guess I was a little unclear in one of my comments, here – sorry about that. *NO* system should have a non-binary goal. No goal is fine (toy), but once you add a goal, it has to be binary.

    When I said the thing about contests having a “measurement” aspect to them, that measurement still has to have a binary judgment on it. So I lifted 300 pounds, you lifted 301, you won the weightlifting contest.

    Hope that makes it clear. No system should have a non-binary goal. It’s like a logically absurd concept to begin with, created because I think we all kind of intrinsically understand that there are values to goals, but making binary goals (especially in single player applications) requires a shitload of balancing work that the “founding fathers” of videogames didn’t have the time or understanding to take on. So we had this weirdo suggestion-goal thing, and players would either play with it like a toy, or house rule it so that it could be played competitively.

  • Jesse Fortner

    Can a non-binary goal such as “get the highest score” be made non-absurd it’s analyzed as a series of smaller goals? Example, in Tetris, once you’ve surpassed your previous high score, each time you score a point, a new goal is created – score another. So you have a series of binary goals that change quickly enough that the player doesn’t think of them as such.

    My concern here is that when confronted with a seeming absurdity like a non-binary goal, the solution is to just throw away the whole concept rather than try to figure out why it seems to work.

  • Jimmy Mac

    Maybe, but does it matter if “get the highest score I can” is a sub-goal on the way to getting on the board? Maybe like “learn how to shoot a fireball” in Street Fighter? The game is giving you feedback about how well you are progressing in the system, but not expecting you to stop. I’m just kind of playing with the ideas, and even if you think about it this way, I’m not sure it changes being a contest not a game in Burgun’s world.

  • Jimmy Mac

    I don’t think everyone agrees with (2) though. A lot of people don’t care about things being empirically true as long as they are “true for me”. For example, Charles Pratt says developing rulesets isn’t his only goal (as opposed to caring about motivation), and games don’t exist unless they are played. Those are statements that privilege subjectivity, and are anti-empirical. He says,”We can make these machines but we can never fully determine how they are used.” If a thing only exists in the mind of a player who is playing it, how can we objectively study it? I think he is saying we should care more about motivations (the subjective experiences of players) rather than goals (the objective, provable features in a game).

    Pratt, I would think, would agree that the *process* of design can be refined, but not designs themselves. It seems like you think that the games themselves can be studied, and that some designs are better than others. The hard version of verificationsim might be wrong, but it seems like you agree with a softer version, that the best evidence is objective and empirical, and that statements like “it’s true for me” would be hard for you to accept. You also reject improvisation as a useful method of design. I think Pratt would say something like, you learn about the formal aspects of design, but when you actually design something, you need to put yourself in an unstructured place to sort of “commune” with the minds of the players who will play your game. You, on the other hand, think that’s pretty much touchy-feely BS and that making considered and thoughtful design is best.

    Don’t mean this to be confrontation, I’m not sure how it comes across in writing, but I’m just genuinely curious about what you think. Pushing back a bit helps me refine my own thinking.

  • Well, sorry, I should have said, anyone worth talking to at least professes to believe (2). Whether all of their beliefs actually reflect an acceptance of (2) is another matter. But I think when asked, most people would agree with the statement.

    Yeah, I think it’s fair to say that the best pursuit of truth lies somewhere in between verificationism – the idea that nothing can be said unless it can be scientifically proven – and post-modernism, which basically is an authoritarian idea that something is basically as true as it can be because someone said it.

    The reality is, of course it’s preferable if we can scientifically prove things, but there are many situations – design is one of them, morality another, economics and soft sciences – wherein we can’t give a lot of hard scientific proofs, but we just have to make judgment calls, and we know enough to know that not all answers are equal. So, we can use reason to make better predictions and determinations. At a later time, it may become possible to scientifically prove aesthetics, and at that time, we will use that hard science instead of the current soft ones.

  • Jereshroom

    That doesn’t really work, because then
    1. most games won’t be nearly as fun and
    2. you have to pretend you don’t know about the goal-creation system.

    Under your proposed system, applied to a game with levels, players start with the goal “beat level one”. So players shouldn’t care whether they beat the level with all their health or one health point. Once they have beaten that level, a new goal is created, “beat level two”. Now, players suddenly do care about how much health they ended the last level with.
    Basically, planning ahead goes from a strategy to a basically a form of cheating (sort of like gaining a reputation of truth-telling in a bluffing game — it hurts you this game but helps you in future games, which shouldn’t matter to you at that time).

  • What you’re asking is “can you prescribe new rules to turn a non binary goal into a binary goal?”. The answer is yes. You still have the awkward strange situation in that game where the game has NOT ENDED but you have already won. What are you doing in those moments, it’s very bizarre.

  • Jesse Fortner

    I’m still talking about description rather than prescription. You suggested, for example that if the goal is only to beat your previous high score, then a player should get 1 point higher, and then quit so as not to make it more difficult the next time. I’ve never actually heard of a player thinking this way, which makes me think that the goal isn’t being analyzed property.

  • No, you’re right, in practice, players don’t think that way. They kind of play with it like a toy, they maybe prescribe, on the fly, some NEW even higher goal (I wonder if I can get to 11,000!!!).

    I only brought up the “technically they should get 1 point higher” as a way of demonstrating the weirdness of this situation when looked at closely, not to suggest that people actually do that.

  • Jereshroom

    I think that works — basically, all of your playing is practice for the true game of which the goal is “get the highest score”. Probably not technically a game by Burgun’s definition, but it seems mostly unproblematic.

  • tinytouchtales

    Great episode, i really liked the part about goals. The question that i’m asking me know is: Should the goal of the game optimally also be the end of the game condition? In the Golf example the completing the 18th hole is the goal and also the game end condition. In Soccer the goal is to score and the time constraint is the end condition. Which is almost identical to a time boxed Highscore game.

  • Yeah, the goal and the end-game condition need to happen at the same time. Although in soccer, I would say it’s actually “at the end game point, have more points than your opponent”. It’s sloppy and has issues but technically it’s fine.